/Friend-Blend-DIP

Primary LanguageJupyter Notebook

Friend Blend

Brief Description

This project contains a program to combine images of two people taken at the same location with similar perspectives to obtain a single image that contains both people. It is achieved using python with the help of opencv.

Documentation of the project can be found here.

Presentation for the final evals can be found here

Setup and Installation

Preliminary setup

  • Install python 3 (version $\geq$ 3.9) from here

Project setup

  • Clone and enter into this repository with the command
git clone https://github.com/Digital-Image-Processing-IIITH/dip-m22-project-big-dipper-22.git
cd dip-m22-project-big-dipper-22

  • Download the test images and outputs from here and extract it to the data directory in the cloned repository

  • Create and activate a python virtual environment inside the repository using

python -m venv .venv
./.venv/scripts/activate
  • To download all the required python modules, run
pip install -r requirements.txt
  • The project code is in the notebook /src/full_new.ipynb
  • Replace the variable image_num with numbers 1-15 in the notebook to run any desired image.
  • The output will be shown at the end of the notebook and is also exported to data/outs/image_num.jpg

Sample outputs

Successful

Inputs:

Image 1: Image 1 Image 2: Image 2

Output: Output

Unsuccessful

Inputs: Image 1: Image 1 Image 2: Image 2

Output: Output Perspective warping fails in the above case since the background does not have enough texture.

Team Information

Team name: Big Dipper Team members:

  • Maulesh Gandhi - 2020112009
  • Sankeerthana Venugopal - 2020102008
  • Sreenya Chitluri - 2020102065
  • Tejah S S - 2020112028

Biblography